Preventive Medicine 69 (2014) 8–12
Contents lists available at ScienceDirect
Preventive Medicine journal homepage: www.elsevier.com/locate/ypmed
Brief Original Report
Non-melanoma skin cancer and NSAID use in women with a history of skin cancer in the Women's Health Initiative☆ Ashley Wysong a,b,⁎,1, Mina S. Ally b,1, Christina S. Gamba c, Manisha Desai c, Susan M. Swetter b,c,d, Kristina Seiffert-Sinha e, Animesh A. Sinha e, Marcia L. Stefanick c, Jean Y. Tang b,c a
University of Southern California, Keck School of Medicine, Department of Dermatology, Los Angeles, CA, United States Stanford University, Department of Dermatology, Redwood City, CA, United States Stanford University School of Medicine and Cancer Institute, Stanford, CA, United States d VA Palo Alto Health Care System, Palo Alto, CA, United States e Department of Dermatology, University at Buffalo, Buffalo, NY, United States b c
a r t i c l e
i n f o
Available online 20 August 2014 Keywords: Non-steroidal Anti-inflammatory drugs Aspirin Non-melanoma skin cancer Female
a b s t r a c t Objective: Evidence for the effect of non-steroidal anti-inflammatory drugs (NSAIDs) on non-melanoma skin cancer (NMSC) risk is inconsistent. We prospectively examined whether regular, inconsistent, or no/low-use of NSAIDs is associated with lower NMSC risk among 54,728 postmenopausal Caucasian women in the Women's Health Initiative Observational Study enrolled between 1993 and 1998. Methods: Logistic regression models were used to assess odds of NMSC after adjusting for skin type, sun exposure history and indication for NSAID use. Results: There were 7652 incident cases of NMSC (median follow-up: 6.9 years). There was no association between regular NSAID-use and NMSC risk relative to no/low-users. However, in a subgroup analysis of 5325 women with a history of skin cancer (incident NMSC: 1897), odds of NMSC were lower among regular NSAID users whether b 5 years (OR 0.82, 95% CI: 0.70–0.95) or ≥5 years (OR 0.82, 95% CI: 0.69–0.98) of use compared to no/low-users. Inconsistent NSAID use and acetaminophen use were not associated with NMSC risk. Conclusion: Overall, NSAID use was not associated with NMSC risk. However, in women with a history of skin cancer, regular NSAID use was associated with 18% lower odds of NMSC. Future studies on potential chemopreventative effects of NSAIDs should focus on subjects with prior history of NMSC. © 2014 Elsevier Inc. All rights reserved.
Introduction Non-melanoma skin cancers (NMSCs) are the most prevalent type of cancer in the US with 20% of white Americans developing NMSC by the age of 70 (Stern, 2010). While mortality from NMSC is low, it is associated with significant morbidity and cost. NMSC care accounts for an estimated annual expenditure of $460 million in the Medicare population alone (John Chen et al., 2006). This has prompted investigations into prevention strategies, including chemoprevention in individuals at increased risk for NMSC. Abbreviations: NSAID, non-steroidal anti-inflammatory drug; ASA, aspirin; NMSC, non-melanoma skin cancer. ☆ Funding: This work was supported by the Women's Health Initiative Program, which is funded by the National Heart, Lung, and Blood Institute, National Institutes of Health, U.S. Department of Health and Human Services through contracts HHSN268201100046C, HHSN268201100001C, HHSN268201100002C, HHSN268201100003C, HHSN268201100004C, and HHSN271201100004C. ⁎ Corresponding author at: Stanford University, Department of Dermatology, 450 Broadway, Pavilion C, Redwood City, CA 94305, United States. Fax: +1 650 496 2573. E-mail address: [email protected] (A. Wysong). 1 Co-first authors.
http://dx.doi.org/10.1016/j.ypmed.2014.08.024 0091-7435/© 2014 Elsevier Inc. All rights reserved.
Cyclooxygenase (COX)-2 enzymes are indicators of inflammation and are upregulated in several cancers (Chan et al., 2011), including NMSC (Higashi et al., 2000). Non-steroidal anti-inflammatory drugs (NSAIDs), which inhibit COX enzymes, are associated with decreased risk of a variety of cancers, including colorectal cancer (Chan et al., 2011) and melanoma (Gamba et al., 2013). Evidence from studies investigating the association of NSAID use and risk of NMSC development is inconsistent. Some animal studies, small clinical trials (Tang et al., 2010) and case–control studies (Torti et al., 2011) have found a protective effect of NSAIDs in NMSC. In contrast, a case–control study (Asgari et al., 2010) and a prospective study (Jeter et al., 2012) showed no significant effect. Due to the limited and contradictory results in the literature, we prospectively examined the association between NSAID use and NMSC in the Women's Health Initiative (WHI) Observational Study (OS) cohort. Methods The design, eligibility criteria, and recruitment methods of WHI OS (NCT00000611) have been previously described (National Institutes of Health
A. Wysong et al. / Preventive Medicine 69 (2014) 8–12
9
Table 1 Baseline characteristics of cohort by NSAID user groupa (aspirin, non-selective non-ASA NSAID, COX-2 inhibitor). Baseline characteristic
Total (N)
Regular use (≥5 years)
Regular use (b5 years)
Inconsistent use
No/low use
N Percent (%) Age (%) 50–59 60–69 70–79 Education (%) High school Some college College BMI (%) b25 25–b30 30+ Smoking (%) Never Past Current Vitamin D intake (diet + supplements) (%) b200 IU/day 200–b400 IU/day 400–b600 IU/day ≥600 IU/day Langleys of exposurec (%) 300–325 350 375–380 400–430 475–500 Physical activity in MET quintile (%) 0–2.25 2.25–8.33 8.33–17.75 17.75+ History of skin cancer (NMSC) (%)b No Yes History of melanoma (%)b No Yes Last medical visit within one year (%) No Yes Oral contraceptive use (%) No Yes Hormone therapy use (%) No Yes Cardiovascular disease ever (%) No Yes Arthritis ever (%) No Yes Migraine ever (%) No Yes NMSC during follow-up (after year 3) (%) No Yes Death during follow-up (after year 3) (%) No Yes Skin reaction to the sun (%) Does not burn Burns, then tans Burns, then tans a minimal amount Burns but does not tan Average time spent outdoors daily in summer as a child (%) Less than 30 min 30 min to 2 h
54,728
7058 12.9
10,795 19.7
8281 15.1
28,594 52.2
17,398 24,753 12,577
24.6 47.7 27.8
26.0 46.6 27.5
35.0 44.9 20.2
34.8 44.2 20.9
10,358 19,708 24,662
19.4 36.2 44.4
20.1 37.9 42.0
19.4 37.2 43.4
18.2 34.9 46.9
23,968 18,585 12,175
38.3 36.2 25.5
40.5 34.5 25.1
37.4 35.0 27.6
48.3 32.9 18.9
27,543 24,272 2913
48.7 46.3 5.1
49.7 45.1 5.2
49.4 45.3 5.3
51.2 43.3 5.5
10,433 19,411 12,626 12,258
17.2 34.5 23.5 24.8
17.1 35.6 23.3 23.9
18.7 36.1 23.4 21.9
20.4 35.5 22.8 21.4
17,727 11,490 6084 8774 10,653
38.4 21.4 10.5 14.5 15.3
34.9 19.3 10.9 16.6 18.3
29.1 20.1 12.9 18.5 19.4
30.9 21.8 10.8 15.5 21.0
16,557 15,136 12,452 10,583
28.4 29.0 23.3 19.3
28.9 27.7 23.4 20.1
28.2 27.2 23.4 21.2
31.8 27.5 22.2 18.5
49,963 4765
90.5 9.5
91.4 8.6
91.5 8.5
91.4 8.6
54,005 723
98.9 1.1
98.4 1.6
98.8 1.2
98.7 1.3
8582 46,146
12.4 87.9
12.9 87.1
14.2 85.8
18.0 82.0
31,337 23,391
58.1 42.0
58.9 41.2
53.4 46.7
57.6 42.4
20,354 34,374
33.3 66.8
35.0 65.0
32.3 67.7
40.4 59.6
44,694 10,034
71.1 29.0
76.6 23.4
84.5 15.5
85.4 14.6
28,714 26,014
42.6 57.5
44.1 55.9
39.2 60.8
61.9 38.1
48,129 6599
85.8 14.2
86.2 13.8
86.0 14.0
89.7 10.3
47,076 7652
85.6 14.4
86.1 13.9
86.3 13.7
86.0 14.0
50,580 4148
91.0 9.0
91.3 8.8
92.5 7.5
93.2 6.8
19,649 14,051 14,876 6152
37.8 24.5 26.8 10.9
36.5 24.5 27.8 11.3
36.1 25.6 26.6 11.7
35.2 26.4 27.2 11.2
1189 14,221
1.9 26.3
2.1 25.5
2.1 25.2
2.3 26.3
P-value
b0.0001
b0.0001
b0.0001
0.0001
b0.0001
b0.0001
b0.0001
0.095
0.052
b0.0001
b0.0001
b0.0001
b0.0001
b0.0001
b0.0001
0.69
b0.0001
0.0001
0.04
(continued on next page)
10
A. Wysong et al. / Preventive Medicine 69 (2014) 8–12
Table 1 (continued) Baseline characteristic
Total (N)
Regular use (≥5 years)
Regular use (b5 years)
Inconsistent use
No/low use
More than 2 h Average time spent outdoors daily in summer now (%) Less than 30 min 30 min to 2 h More than 2 h Sunscreen SPF (%) No sunscreen SPF 2–14 SPF 15–25 SPF 25+
39,318
71.9
72.4
72.7
71.4
16,373 27,672 10,683
30.9 50.8 18.3
31.3 50.3 18.4
30.5 51.1 18.3
29.0 50.5 20.6
25,860 2687 16,740 9441
48.2 4.5 30.8 16.5
47.7 4.7 30.3 17.4
46.3 5.1 31.0 17.6
47.1 5.1 30.5 17.3
P-value b0.0001
0.128
ASA — aspirin; COX — cyclooxygenase; NMSC — non-melanoma skin cancer. Percentages refer to the proportion of women in the user group column. a No/low use was defined as participants who reported NSAID use less than twice weekly for at least 2 weeks preceding the questionnaire at baseline AND year 3. Regular users were participants who reported NSAID use ≥2 times/week for at least 2 weeks preceding the questionnaire at baseline AND year 3. Inconsistent users reported NSAID use ≥2 times/week at baseline OR year 3. Duration of NSAID use was recorded at baseline and categorized as b5 years or ≥5 years among regular users. b Personal history of skin cancer (163 women had a personal history of both NMSC and melanoma).
et al., 1998; Hays et al., 2003). Briefly, the OS enrolled 93,676 postmenopausal women (50–79 years) between 1993 and 1998 at 40 sites across the U.S. Most were initially screened for the randomized trials and were ineligible or unwilling to participate, but were eligible for the OS. We limited our cohort to immunocompetent Caucasian participants who had complete information for all variables. Approval was obtained from the Institutional Review Boards at each participating institution. Demographics, medical history, diet and supplement use, physical activity, smoking status, and physical activity measures were obtained by questionnaire, interview or physical exam at baseline. Information on average daily time spent in the sun, skin reaction to the sun and sunscreen use in the past year was collected by questionnaire. Baseline AND year 3 medications were collected using an interviewadministered questionnaire (Form 44-Current Medications; accessed at cleo. whi.org). Women were considered NSAID-users if they reported NSAID use at least twice/week for at least 2 weeks preceding the interview as previously classified (Gamba et al., 2013). Information on NSAID type (e.g. aspirin, ibuprofen), strength and duration of use was collected. Four mutually exclusive NSAID exposure groups were defined to avoid multiple comparisons and to allow for a single, four-way comparison: no/low NSAID use, inconsistent use,
regular use for b 5 years and regular use for ≥5 years. Regular use was defined as NSAID use ≥2 times/week for at least 2 weeks preceding the questionnaire at baseline AND year 3 (Table 1). The primary outcome was self-reported NMSC in an annual questionnaire after year 3. Information on NMSC subtype (BCC vs. SCC) and ICD codes were not available. Statistical analyses Logistic regression analyses were used to address whether risk of NMSC differs among the exposure groups, adjusting for the following potential confounders: age, regional solar radiation (Langleys), education, body mass index, smoking status, vitamin D use, physical activity, personal history of NMSC, personal history of melanoma, skin reaction to the sun, current and childhood summer sun exposure, sunscreen use, time since last medical visit, history of cardiovascular disease, arthritis and migraines, and any other NSAID use. Confounders were modeled categorically when possible. We performed the above analyses for acetaminophen use (non-NSAID) to identify residual confounding by the use of any pain medication. We also used F-tests to assess whether 8 pre-specified known and potential baseline skin cancer risk factors (age, history of skin cancer, smoking, childhood and current summer-sun
Table 2 Results from logistic regression models for NMSC during follow-up from year 3 to closeout. Variable
Odds ratio Any NSAID use Regular, ≥5 years Regular, b5 years Inconsistent No/low use Aspirin use only Regular, ≥5 years Regular, b5 years Inconsistent No/low use Non-selective NSAID use only Regular, ≥5 years Regular, b5 years Inconsistent No/low use Acetaminophen use Regular, ≥5 years Regular, b5 years Inconsistent No/low use
Fully adjusted modela
Age adjusted model P-value
Odds ratio
0.48 0.99 (0.92, 1.07) 0.95 (0.89, 1.02) 0.97 (0.90, 1.04) Ref
0.60 0.99 (0.91, 1.07) 0.96 (0.89, 1.02) 0.98 (0.91, 1.05) Ref
0.67 1.01 (0.93, 1.10) 0.99 (0.89, 1.09) 0.97 (0.91, 1.03) Ref
0.56 1.01 (0.92, 1.10) 0.98 (0.89, 1.09) 0.96 (0.90, 1.02) Ref
0.84 0.97 (0.86, 1.10) 0.98 (0.87, 1.11) 0.97 (0.92, 1.04) Ref
0.82 0.96 (0.84, 1.09) 1.01 (0.88, 1.14) 0.98 (0.91, 1.04) Ref
0.46 1.04 (0.90, 1.20) 1.04 (0.87, 1.23) 1.05 (0.99, 1.13) Ref
P-value
0.26 1.07 (0.92, 1.24) 1.08 (0.90, 1.29) 1.07 (0.99, 1.14) Ref
NMSC — non-melanoma skin cancer; NSAID — non-steroidal anti-inflammatory drug. a Adjusted for the following covariates that are related to NSAID use or risk factors for skin cancer: age, regional solar radiation (Langleys), education, body mass index (BMI), smoking status, vitamin D use, physical activity, personal history of NMSC, personal history of melanoma, skin reaction to the sun, current and childhood summer sun exposure, sunscreen use, time since last medical visit, history of cardiovascular disease, arthritis and migraines, and any other NSAID use. All statistical analyses were performed using SAS 9.2 (SAS Institute Inc., Cary, NC). Statistical tests were all 2 sided and were analyzed at the 0.05 level of significance.
A. Wysong et al. / Preventive Medicine 69 (2014) 8–12 Table 3 Odds ratios of NMSC according to the use of any NSAID stratified by the effect modifier of history of skin cancer (melanoma and NMSC). P = 0.04 for the interaction between history of skin cancer and NSAID use.
No history of skin cancer (N = 49,403) Regular use of any NSAID ≥5 years Regular use of any NSAID b5 years Inconsistent use of any NSAID History of skin cancer (N = 5325) Regular use of any NSAID ≥5 years Regular use of any NSAID b5 years Inconsistent use of any NSAID
Age adjusted
Fully adjusted
Odds ratio (95% CI)a
Odds ratio (95% CI)a
1.03 (0.95, 1.12) 1.00 (0.93, 1.07) 1.00 (0.92, 1.09)
1.03 (0.95, 1.13) 0.99 (0.92, 1.06) 1.00 (0.92, 1.08)
0.84 (0.71, 1.00) 0.83 (0.71, 0.96) 0.90 (0.76, 1.06)
0.82 (0.69, 0.98) 0.82 (0.70, 0.95) 0.90 (0.76, 1.07)
Fully adjusted for the following covariates: age, regional solar radiation (Langleys), education, body mass index (BMI), smoking status, vitamin D use, physical activity, skin reaction to the sun, current and childhood summer sun exposure, sunscreen use, time since last medical visit, history of cardiovascular disease, arthritis and migraines, and any other NSAID use. a Versus no/low use.
exposure, regional solar radiation [Langley's], vitamin D intake, BMI) or history of cardiovascular disease (CVD) (as a medical indication for NSAID use) could, via interaction, modify the association of NSAID use on the odds of developing NMSC. These nine pre-specified variables were categorically modeled as effect modifiers. All statistical analyses were performed using SAS 9.2 (SAS Institute Inc., Cary, NC).
Results A total of 54,728 women were eligible for inclusion in this study. Participants had a mean age of 63.5 years and mean duration of 6.7 years of NSAID use. Of 17, 853 regular NSAID-users, 47% (8411) used ASA and the remainder used non-ASA NSAIDs. Of the participants who reported regular ASA use, 2059 (24%) reported low-dose ASA (81 mg) use. Among the non-ASA NSAID users, the most commonly reported medications at baseline included in descending order ibuprofen, naproxen, nabumetone and diclofenac. Statistically significant differences between NSAID exposure groups were noted in several baseline factors reflecting the effects of the large WHI sample size on small differences in percentages (Table 1). Overall, regular NSAID-users were older and more likely to report a history of CVD compared with inconsistent or no/low NSAID-users. During a median follow-up time of 6.9 years, there were 7652 incident NMSC cases. No significant differences in incident NMSC were seen between the 4 exposure groups (Table 1). There was no association between regular NSAID use, type of NSAID use, duration of use and the risk of NMSC relative to no/lowusers in either age-adjusted or fully-adjusted models (Table 2). There was no association between acetaminophen use and risk of NMSC. The association between regular NSAID use and risk of NMSC did not differ by known and potential skin cancer risk factors such as age (P =0.08), BMI (P = 0.79), smoking status (P = 0.36), current sun exposure (P = 0.23), childhood summer sun exposure (P = 0.71), regional solar radiation (P = 0.90), vitamin D intake (P = 0.55) or history of CVD (P = 0.71) at baseline in pre-specified subgroup analyses (effect modifiers). However, the association between NSAID use and risk of NMSC differed based on history of skin cancer (P = 0.05). In a priori subset analysis of 5325 women who reported a history of skin cancer at baseline, there was a reduced odds of NMSC among regular NSAID users for ≥ 5 years (OR 0.82, 95% CI: 0.69–0.98) and b 5 years (OR 0.82, 95% CI: 0.70–0.95) when compared to no/low-users (Pinteraction = 0.04) (Table 3). Inconsistent NSAID use was not associated with statistically significant reduced odds of NMSC (OR 0.90, 95% CI: 0.76–1.07).
11
Discussion Overall, there was no association between NSAID use and NMSC risk in 54,728 post-menopausal women. However, in 5325 women with prior skin cancer history, regular NSAID use was associated with significantly decreased odds of NMSC by 18% relative to no/low-users. To our knowledge, the WHI OS is the first study to demonstrate the association between regular NSAID use and NMSC in women with a history of skin cancer. Our results are consistent with two case–control studies (Johannesdottir et al., 2012; Torti et al., 2011) and one randomized-controlled trial (Elmets et al., 2010). A double-blind, placebo-controlled trial of 240 subjects with actinic keratosis showed a 59% reduction in NMSC in the celecoxib arm versus placebo after 11 months (Elmets et al., 2010), suggesting that our follow-up of 6.9 years is sufficient to observe the preventative effect of NSAIDs on NMSC. In a Denmark population-based case–control study, men and women with NSAID ever-use (N2 prescriptions) compared with nonuse (b2 prescriptions) had a decreased risk of NMSC, especially for longterm (N7 years) and high-intensity use (Johannesdottir et al., 2012). Another case–control study found that regular use of full dose (N200 mg aspirin), but not low dose (b200 mg) NSAIDs was associated reduced odds of NMSC (Butler et al., 2005). Interestingly, we did not see a significant dose–response relationship with longer duration of NSAID use. One explanation for our results may be differences in study design and definitions of “regular” NSAID use. Further studies are needed to evaluate the ideal duration and dosage of NSAID use for NMSC chemoprevention as this has implications for limiting side effects from long-term use. Our findings differed from those of a case–control study (Asgari et al., 2010) and cohort studies (Friis et al., 2003; Grau et al., 2006; Jeter et al., 2012). The Nurses' Health Study cohort found no association between NSAIDs and NMSC risk; however, women with a history of skin cancer were excluded (Jeter et al., 2012). Likewise a Kaiser case–control study showed no reduction in risk but also excluded subjects with prior skin cancer history in their controls (Asgari et al., 2010). Other explanations for the inconsistency with our findings include the evaluation of NSAIDs on the risk of SCC alone rather than all NMSCs (Asgari et al., 2010), smaller sample size (Asgari et al., 2010; Grau et al., 2006), differing definitions of regular NSAID use (Asgari et al., 2010; Grau et al., 2006; Jeter et al., 2012) and exposure measurements relying strictly on recall without medication validation (Jeter et al., 2012). We believe our findings that NSAIDs are associated with decreased odds of NMSC only in women with a prior history of skin cancer are consistent with studies of other malignancies in high-risk populations (Ferrandez et al., 2012). Specifically, chronic inflammation correlates with increased risk of developing cancer in the affected organ and can be caused by ultraviolet radiation. Subjects with prior skin cancer have had greater ultraviolet-B exposure (Chang et al., 2010) and have higher epidermal COX2 expression and NF-kB activation, resulting in inflammation and cancer pathogenesis. These compounds are inhibited by NSAIDs, making this population more amenable to chemoprevention (Yamamoto and Gaynor, 2001). Our findings suggest that NSAIDs may be more effective in secondary prevention among high-risk patients rather than in primary prevention of NMSC, much like studies showing greater benefits of antihypertensives in those with higher versus lower categories of hypertension (Arguedas et al., 2009). The strengths of our study include the prospective design, large number of cases and long follow-up period as well as detailed information on key skin cancer risk factors. In addition, NSAID use was validated with pill bottle labels and/or prescription records, which may not have been done in other studies (Jeter et al., 2012). Limitations include selfreporting of NMSC (without medical record adjudication), although the validity of patient self-reporting of skin cancer is high (Ming et al., 2004) and lack of information on NMSC subtype (BCC vs SCC), which may have yielded information on whether NSAIDs have a specific effect based on subtype, as demonstrated previously (Johannesdottir et al.,
12
A. Wysong et al. / Preventive Medicine 69 (2014) 8–12
2012; Torti et al., 2011). Finally, we did not stratify the analysis by NSAID type or dosage among women with a history of skin cancer and cannot comment on whether ASA versus non-ASA NSAIDS or high versus low dosage NSAIDs were driving the protective effect. Conclusion In conclusion, our results indicate that although NSAID use was not associated with NMSC risk overall, postmenopausal Caucasian women with a history of skin cancer who regularly used NSAIDs had significantly lower odds of future NMSC. Future studies on potential chemopreventative effects of NSAIDs should be focused on subjects with prior history of NMSC. Conflict of Interest statement The authors declare that they have no conflict of interest.
Acknowledgments The authors would like to thank Kristopher Kapphahn and Haley Hedlin for their help with the statistical analysis. References Arguedas, J.A., Perez, M.I., Wright, J.M., 2009. Treatment blood pressure targets for hypertension. Cochrane Database Syst. Rev. CD004349 (Online). Asgari, M.M., Chren, M.M., Warton, E.M., Friedman, G.D., White, E., 2010. Association between nonsteroidal anti-inflammatory drug use and cutaneous squamous cell carcinoma. Arch. Dermatol. 146, 388–395. Butler, G.J., Neale, R., Green, A.C., Pandeya, N., Whiteman, D.C., 2005. Nonsteroidal anti-inflammatory drugs and the risk of actinic keratoses and squamous cell cancers of the skin. J. Am. Acad. Dermatol. 53, 966–972. Chan, A.T., Ogino, S., Giovannucci, E.L., Fuchs, C.S., 2011. Inflammatory markers are associated with risk of colorectal cancer and chemopreventive response to anti-inflammatory drugs. Gastroenterology 140, 799–808 (quiz e11). Chang, N.B., Feng, R., Gao, Z., Gao, W., 2010. Skin cancer incidence is highly associated with ultraviolet-B radiation history. Int. J. Hyg. Environ. Health 213, 359–368.
Elmets, C.A., Viner, J.L., Pentland, A.P., et al., 2010. Chemoprevention of nonmelanoma skin cancer with celecoxib: a randomized, double-blind, placebo-controlled trial. J. Natl. Cancer Inst. 102, 1835–1844. Ferrandez, A., Piazuelo, E., Castells, A., 2012. Aspirin and the prevention of colorectal cancer. Best Pract. Res. Clin. Gastroenterol. 26, 185–195. Friis, S., Sorensen, H.T., McLaughlin, J.K., Johnsen, S.P., Blot, W.J., Olsen, J.H., 2003. A population-based cohort study of the risk of colorectal and other cancers among users of low-dose aspirin. Br. J. Cancer 88, 684–688. Gamba, C.A., Swetter, S.M., Stefanick, M.L., et al., 2013. Aspirin is associated with lower melanoma risk among postmenopausal Caucasian women: the Women's Health Initiative. Cancer 119 (8), 1562–1569. Grau, M.V., Baron, J.A., Langholz, B., et al., 2006. Effect of NSAIDs on the recurrence of nonmelanoma skin cancer. Int. J. Cancer 119, 682–686. Hays, J., Hunt, J.R., Hubbell, F.A., et al., 2003. The Women's Health Initiative recruitment methods and results. Ann. Epidemiol. 13, S18–S77. Higashi, Y., Kanekura, T., Kanzaki, T., 2000. Enhanced expression of cyclooxygenase (COX)-2 in human skin epidermal cancer cells: evidence for growth suppression by inhibiting COX-2 expression. Int. J. Cancer 86, 667–671. Jeter, J.M., Han, J., Martinez, M.E., Alberts, D.S., Qureshi, A.A., Feskanich, D., 2012. Non-steroidal anti-inflammatory drugs, acetaminophen, and risk of skin cancer in the Nurses' Health Study. Cancer Causes Control 23, 1451–1461. Johannesdottir, S.A., Chang, E.T., Mehnert, F., Schmidt, M., Olesen, A.B., Sorensen, H.T., 2012. Nonsteroidal anti-inflammatory drugs and the risk of skin cancer: a population-based case–control study. Cancer 118, 4768–4776. John Chen, G., Yelverton, C.B., Polisetty, S.S., et al., 2006. Treatment patterns and cost of nonmelanoma skin cancer management. Dermatol. Surg. 32, 1266–1271. Ming, M.E., Levy, R.M., Hoffstad, O.J., Filip, J., Gimotty, P.A., Margolis, D.J., 2004. Validity of patient self-reported history of skin cancer. Arch. Dermatol. 140, 730–735. National Institutes of Health, Women's Health Initiative, Form 44-Current Medications, 1998. Design of the Women's Health Initiative clinical trial and observational study. The Women's Health Initiative Study Group. Control. Clin. Trials 19, 61–109. Stern, R.S., 2010. Prevalence of a history of skin cancer in 2007: results of an incidence-based model. Arch. Dermatol. 146, 279–282. Tang, J.Y., Aszterbaum, M., Athar, M., et al., 2010. Basal cell carcinoma chemoprevention with nonsteroidal anti-inflammatory drugs in genetically predisposed PTCH1+/− humans and mice. Cancer Prev. Res. (Phila.) 3, 25–34. Torti, D.C., Christensen, B.C., Storm, C.A., et al., 2011. Analgesic and nonsteroidal anti-inflammatory use in relation to nonmelanoma skin cancer: a population-based case–control study. J. Am. Acad. Dermatol. 65, 304–312. Yamamoto, Y., Gaynor, R.B., 2001. Therapeutic potential of inhibition of the NF-kappaB pathway in the treatment of inflammation and cancer. J. Clin. Invest. 107, 135–142.
Contents lists available at ScienceDirect
Preventive Medicine journal homepage: www.elsevier.com/locate/ypmed
Brief Original Report
Non-melanoma skin cancer and NSAID use in women with a history of skin cancer in the Women's Health Initiative☆ Ashley Wysong a,b,⁎,1, Mina S. Ally b,1, Christina S. Gamba c, Manisha Desai c, Susan M. Swetter b,c,d, Kristina Seiffert-Sinha e, Animesh A. Sinha e, Marcia L. Stefanick c, Jean Y. Tang b,c a
University of Southern California, Keck School of Medicine, Department of Dermatology, Los Angeles, CA, United States Stanford University, Department of Dermatology, Redwood City, CA, United States Stanford University School of Medicine and Cancer Institute, Stanford, CA, United States d VA Palo Alto Health Care System, Palo Alto, CA, United States e Department of Dermatology, University at Buffalo, Buffalo, NY, United States b c
a r t i c l e
i n f o
Available online 20 August 2014 Keywords: Non-steroidal Anti-inflammatory drugs Aspirin Non-melanoma skin cancer Female
a b s t r a c t Objective: Evidence for the effect of non-steroidal anti-inflammatory drugs (NSAIDs) on non-melanoma skin cancer (NMSC) risk is inconsistent. We prospectively examined whether regular, inconsistent, or no/low-use of NSAIDs is associated with lower NMSC risk among 54,728 postmenopausal Caucasian women in the Women's Health Initiative Observational Study enrolled between 1993 and 1998. Methods: Logistic regression models were used to assess odds of NMSC after adjusting for skin type, sun exposure history and indication for NSAID use. Results: There were 7652 incident cases of NMSC (median follow-up: 6.9 years). There was no association between regular NSAID-use and NMSC risk relative to no/low-users. However, in a subgroup analysis of 5325 women with a history of skin cancer (incident NMSC: 1897), odds of NMSC were lower among regular NSAID users whether b 5 years (OR 0.82, 95% CI: 0.70–0.95) or ≥5 years (OR 0.82, 95% CI: 0.69–0.98) of use compared to no/low-users. Inconsistent NSAID use and acetaminophen use were not associated with NMSC risk. Conclusion: Overall, NSAID use was not associated with NMSC risk. However, in women with a history of skin cancer, regular NSAID use was associated with 18% lower odds of NMSC. Future studies on potential chemopreventative effects of NSAIDs should focus on subjects with prior history of NMSC. © 2014 Elsevier Inc. All rights reserved.
Introduction Non-melanoma skin cancers (NMSCs) are the most prevalent type of cancer in the US with 20% of white Americans developing NMSC by the age of 70 (Stern, 2010). While mortality from NMSC is low, it is associated with significant morbidity and cost. NMSC care accounts for an estimated annual expenditure of $460 million in the Medicare population alone (John Chen et al., 2006). This has prompted investigations into prevention strategies, including chemoprevention in individuals at increased risk for NMSC. Abbreviations: NSAID, non-steroidal anti-inflammatory drug; ASA, aspirin; NMSC, non-melanoma skin cancer. ☆ Funding: This work was supported by the Women's Health Initiative Program, which is funded by the National Heart, Lung, and Blood Institute, National Institutes of Health, U.S. Department of Health and Human Services through contracts HHSN268201100046C, HHSN268201100001C, HHSN268201100002C, HHSN268201100003C, HHSN268201100004C, and HHSN271201100004C. ⁎ Corresponding author at: Stanford University, Department of Dermatology, 450 Broadway, Pavilion C, Redwood City, CA 94305, United States. Fax: +1 650 496 2573. E-mail address: [email protected] (A. Wysong). 1 Co-first authors.
http://dx.doi.org/10.1016/j.ypmed.2014.08.024 0091-7435/© 2014 Elsevier Inc. All rights reserved.
Cyclooxygenase (COX)-2 enzymes are indicators of inflammation and are upregulated in several cancers (Chan et al., 2011), including NMSC (Higashi et al., 2000). Non-steroidal anti-inflammatory drugs (NSAIDs), which inhibit COX enzymes, are associated with decreased risk of a variety of cancers, including colorectal cancer (Chan et al., 2011) and melanoma (Gamba et al., 2013). Evidence from studies investigating the association of NSAID use and risk of NMSC development is inconsistent. Some animal studies, small clinical trials (Tang et al., 2010) and case–control studies (Torti et al., 2011) have found a protective effect of NSAIDs in NMSC. In contrast, a case–control study (Asgari et al., 2010) and a prospective study (Jeter et al., 2012) showed no significant effect. Due to the limited and contradictory results in the literature, we prospectively examined the association between NSAID use and NMSC in the Women's Health Initiative (WHI) Observational Study (OS) cohort. Methods The design, eligibility criteria, and recruitment methods of WHI OS (NCT00000611) have been previously described (National Institutes of Health
A. Wysong et al. / Preventive Medicine 69 (2014) 8–12
9
Table 1 Baseline characteristics of cohort by NSAID user groupa (aspirin, non-selective non-ASA NSAID, COX-2 inhibitor). Baseline characteristic
Total (N)
Regular use (≥5 years)
Regular use (b5 years)
Inconsistent use
No/low use
N Percent (%) Age (%) 50–59 60–69 70–79 Education (%) High school Some college College BMI (%) b25 25–b30 30+ Smoking (%) Never Past Current Vitamin D intake (diet + supplements) (%) b200 IU/day 200–b400 IU/day 400–b600 IU/day ≥600 IU/day Langleys of exposurec (%) 300–325 350 375–380 400–430 475–500 Physical activity in MET quintile (%) 0–2.25 2.25–8.33 8.33–17.75 17.75+ History of skin cancer (NMSC) (%)b No Yes History of melanoma (%)b No Yes Last medical visit within one year (%) No Yes Oral contraceptive use (%) No Yes Hormone therapy use (%) No Yes Cardiovascular disease ever (%) No Yes Arthritis ever (%) No Yes Migraine ever (%) No Yes NMSC during follow-up (after year 3) (%) No Yes Death during follow-up (after year 3) (%) No Yes Skin reaction to the sun (%) Does not burn Burns, then tans Burns, then tans a minimal amount Burns but does not tan Average time spent outdoors daily in summer as a child (%) Less than 30 min 30 min to 2 h
54,728
7058 12.9
10,795 19.7
8281 15.1
28,594 52.2
17,398 24,753 12,577
24.6 47.7 27.8
26.0 46.6 27.5
35.0 44.9 20.2
34.8 44.2 20.9
10,358 19,708 24,662
19.4 36.2 44.4
20.1 37.9 42.0
19.4 37.2 43.4
18.2 34.9 46.9
23,968 18,585 12,175
38.3 36.2 25.5
40.5 34.5 25.1
37.4 35.0 27.6
48.3 32.9 18.9
27,543 24,272 2913
48.7 46.3 5.1
49.7 45.1 5.2
49.4 45.3 5.3
51.2 43.3 5.5
10,433 19,411 12,626 12,258
17.2 34.5 23.5 24.8
17.1 35.6 23.3 23.9
18.7 36.1 23.4 21.9
20.4 35.5 22.8 21.4
17,727 11,490 6084 8774 10,653
38.4 21.4 10.5 14.5 15.3
34.9 19.3 10.9 16.6 18.3
29.1 20.1 12.9 18.5 19.4
30.9 21.8 10.8 15.5 21.0
16,557 15,136 12,452 10,583
28.4 29.0 23.3 19.3
28.9 27.7 23.4 20.1
28.2 27.2 23.4 21.2
31.8 27.5 22.2 18.5
49,963 4765
90.5 9.5
91.4 8.6
91.5 8.5
91.4 8.6
54,005 723
98.9 1.1
98.4 1.6
98.8 1.2
98.7 1.3
8582 46,146
12.4 87.9
12.9 87.1
14.2 85.8
18.0 82.0
31,337 23,391
58.1 42.0
58.9 41.2
53.4 46.7
57.6 42.4
20,354 34,374
33.3 66.8
35.0 65.0
32.3 67.7
40.4 59.6
44,694 10,034
71.1 29.0
76.6 23.4
84.5 15.5
85.4 14.6
28,714 26,014
42.6 57.5
44.1 55.9
39.2 60.8
61.9 38.1
48,129 6599
85.8 14.2
86.2 13.8
86.0 14.0
89.7 10.3
47,076 7652
85.6 14.4
86.1 13.9
86.3 13.7
86.0 14.0
50,580 4148
91.0 9.0
91.3 8.8
92.5 7.5
93.2 6.8
19,649 14,051 14,876 6152
37.8 24.5 26.8 10.9
36.5 24.5 27.8 11.3
36.1 25.6 26.6 11.7
35.2 26.4 27.2 11.2
1189 14,221
1.9 26.3
2.1 25.5
2.1 25.2
2.3 26.3
P-value
b0.0001
b0.0001
b0.0001
0.0001
b0.0001
b0.0001
b0.0001
0.095
0.052
b0.0001
b0.0001
b0.0001
b0.0001
b0.0001
b0.0001
0.69
b0.0001
0.0001
0.04
(continued on next page)
10
A. Wysong et al. / Preventive Medicine 69 (2014) 8–12
Table 1 (continued) Baseline characteristic
Total (N)
Regular use (≥5 years)
Regular use (b5 years)
Inconsistent use
No/low use
More than 2 h Average time spent outdoors daily in summer now (%) Less than 30 min 30 min to 2 h More than 2 h Sunscreen SPF (%) No sunscreen SPF 2–14 SPF 15–25 SPF 25+
39,318
71.9
72.4
72.7
71.4
16,373 27,672 10,683
30.9 50.8 18.3
31.3 50.3 18.4
30.5 51.1 18.3
29.0 50.5 20.6
25,860 2687 16,740 9441
48.2 4.5 30.8 16.5
47.7 4.7 30.3 17.4
46.3 5.1 31.0 17.6
47.1 5.1 30.5 17.3
P-value b0.0001
0.128
ASA — aspirin; COX — cyclooxygenase; NMSC — non-melanoma skin cancer. Percentages refer to the proportion of women in the user group column. a No/low use was defined as participants who reported NSAID use less than twice weekly for at least 2 weeks preceding the questionnaire at baseline AND year 3. Regular users were participants who reported NSAID use ≥2 times/week for at least 2 weeks preceding the questionnaire at baseline AND year 3. Inconsistent users reported NSAID use ≥2 times/week at baseline OR year 3. Duration of NSAID use was recorded at baseline and categorized as b5 years or ≥5 years among regular users. b Personal history of skin cancer (163 women had a personal history of both NMSC and melanoma).
et al., 1998; Hays et al., 2003). Briefly, the OS enrolled 93,676 postmenopausal women (50–79 years) between 1993 and 1998 at 40 sites across the U.S. Most were initially screened for the randomized trials and were ineligible or unwilling to participate, but were eligible for the OS. We limited our cohort to immunocompetent Caucasian participants who had complete information for all variables. Approval was obtained from the Institutional Review Boards at each participating institution. Demographics, medical history, diet and supplement use, physical activity, smoking status, and physical activity measures were obtained by questionnaire, interview or physical exam at baseline. Information on average daily time spent in the sun, skin reaction to the sun and sunscreen use in the past year was collected by questionnaire. Baseline AND year 3 medications were collected using an interviewadministered questionnaire (Form 44-Current Medications; accessed at cleo. whi.org). Women were considered NSAID-users if they reported NSAID use at least twice/week for at least 2 weeks preceding the interview as previously classified (Gamba et al., 2013). Information on NSAID type (e.g. aspirin, ibuprofen), strength and duration of use was collected. Four mutually exclusive NSAID exposure groups were defined to avoid multiple comparisons and to allow for a single, four-way comparison: no/low NSAID use, inconsistent use,
regular use for b 5 years and regular use for ≥5 years. Regular use was defined as NSAID use ≥2 times/week for at least 2 weeks preceding the questionnaire at baseline AND year 3 (Table 1). The primary outcome was self-reported NMSC in an annual questionnaire after year 3. Information on NMSC subtype (BCC vs. SCC) and ICD codes were not available. Statistical analyses Logistic regression analyses were used to address whether risk of NMSC differs among the exposure groups, adjusting for the following potential confounders: age, regional solar radiation (Langleys), education, body mass index, smoking status, vitamin D use, physical activity, personal history of NMSC, personal history of melanoma, skin reaction to the sun, current and childhood summer sun exposure, sunscreen use, time since last medical visit, history of cardiovascular disease, arthritis and migraines, and any other NSAID use. Confounders were modeled categorically when possible. We performed the above analyses for acetaminophen use (non-NSAID) to identify residual confounding by the use of any pain medication. We also used F-tests to assess whether 8 pre-specified known and potential baseline skin cancer risk factors (age, history of skin cancer, smoking, childhood and current summer-sun
Table 2 Results from logistic regression models for NMSC during follow-up from year 3 to closeout. Variable
Odds ratio Any NSAID use Regular, ≥5 years Regular, b5 years Inconsistent No/low use Aspirin use only Regular, ≥5 years Regular, b5 years Inconsistent No/low use Non-selective NSAID use only Regular, ≥5 years Regular, b5 years Inconsistent No/low use Acetaminophen use Regular, ≥5 years Regular, b5 years Inconsistent No/low use
Fully adjusted modela
Age adjusted model P-value
Odds ratio
0.48 0.99 (0.92, 1.07) 0.95 (0.89, 1.02) 0.97 (0.90, 1.04) Ref
0.60 0.99 (0.91, 1.07) 0.96 (0.89, 1.02) 0.98 (0.91, 1.05) Ref
0.67 1.01 (0.93, 1.10) 0.99 (0.89, 1.09) 0.97 (0.91, 1.03) Ref
0.56 1.01 (0.92, 1.10) 0.98 (0.89, 1.09) 0.96 (0.90, 1.02) Ref
0.84 0.97 (0.86, 1.10) 0.98 (0.87, 1.11) 0.97 (0.92, 1.04) Ref
0.82 0.96 (0.84, 1.09) 1.01 (0.88, 1.14) 0.98 (0.91, 1.04) Ref
0.46 1.04 (0.90, 1.20) 1.04 (0.87, 1.23) 1.05 (0.99, 1.13) Ref
P-value
0.26 1.07 (0.92, 1.24) 1.08 (0.90, 1.29) 1.07 (0.99, 1.14) Ref
NMSC — non-melanoma skin cancer; NSAID — non-steroidal anti-inflammatory drug. a Adjusted for the following covariates that are related to NSAID use or risk factors for skin cancer: age, regional solar radiation (Langleys), education, body mass index (BMI), smoking status, vitamin D use, physical activity, personal history of NMSC, personal history of melanoma, skin reaction to the sun, current and childhood summer sun exposure, sunscreen use, time since last medical visit, history of cardiovascular disease, arthritis and migraines, and any other NSAID use. All statistical analyses were performed using SAS 9.2 (SAS Institute Inc., Cary, NC). Statistical tests were all 2 sided and were analyzed at the 0.05 level of significance.
A. Wysong et al. / Preventive Medicine 69 (2014) 8–12 Table 3 Odds ratios of NMSC according to the use of any NSAID stratified by the effect modifier of history of skin cancer (melanoma and NMSC). P = 0.04 for the interaction between history of skin cancer and NSAID use.
No history of skin cancer (N = 49,403) Regular use of any NSAID ≥5 years Regular use of any NSAID b5 years Inconsistent use of any NSAID History of skin cancer (N = 5325) Regular use of any NSAID ≥5 years Regular use of any NSAID b5 years Inconsistent use of any NSAID
Age adjusted
Fully adjusted
Odds ratio (95% CI)a
Odds ratio (95% CI)a
1.03 (0.95, 1.12) 1.00 (0.93, 1.07) 1.00 (0.92, 1.09)
1.03 (0.95, 1.13) 0.99 (0.92, 1.06) 1.00 (0.92, 1.08)
0.84 (0.71, 1.00) 0.83 (0.71, 0.96) 0.90 (0.76, 1.06)
0.82 (0.69, 0.98) 0.82 (0.70, 0.95) 0.90 (0.76, 1.07)
Fully adjusted for the following covariates: age, regional solar radiation (Langleys), education, body mass index (BMI), smoking status, vitamin D use, physical activity, skin reaction to the sun, current and childhood summer sun exposure, sunscreen use, time since last medical visit, history of cardiovascular disease, arthritis and migraines, and any other NSAID use. a Versus no/low use.
exposure, regional solar radiation [Langley's], vitamin D intake, BMI) or history of cardiovascular disease (CVD) (as a medical indication for NSAID use) could, via interaction, modify the association of NSAID use on the odds of developing NMSC. These nine pre-specified variables were categorically modeled as effect modifiers. All statistical analyses were performed using SAS 9.2 (SAS Institute Inc., Cary, NC).
Results A total of 54,728 women were eligible for inclusion in this study. Participants had a mean age of 63.5 years and mean duration of 6.7 years of NSAID use. Of 17, 853 regular NSAID-users, 47% (8411) used ASA and the remainder used non-ASA NSAIDs. Of the participants who reported regular ASA use, 2059 (24%) reported low-dose ASA (81 mg) use. Among the non-ASA NSAID users, the most commonly reported medications at baseline included in descending order ibuprofen, naproxen, nabumetone and diclofenac. Statistically significant differences between NSAID exposure groups were noted in several baseline factors reflecting the effects of the large WHI sample size on small differences in percentages (Table 1). Overall, regular NSAID-users were older and more likely to report a history of CVD compared with inconsistent or no/low NSAID-users. During a median follow-up time of 6.9 years, there were 7652 incident NMSC cases. No significant differences in incident NMSC were seen between the 4 exposure groups (Table 1). There was no association between regular NSAID use, type of NSAID use, duration of use and the risk of NMSC relative to no/lowusers in either age-adjusted or fully-adjusted models (Table 2). There was no association between acetaminophen use and risk of NMSC. The association between regular NSAID use and risk of NMSC did not differ by known and potential skin cancer risk factors such as age (P =0.08), BMI (P = 0.79), smoking status (P = 0.36), current sun exposure (P = 0.23), childhood summer sun exposure (P = 0.71), regional solar radiation (P = 0.90), vitamin D intake (P = 0.55) or history of CVD (P = 0.71) at baseline in pre-specified subgroup analyses (effect modifiers). However, the association between NSAID use and risk of NMSC differed based on history of skin cancer (P = 0.05). In a priori subset analysis of 5325 women who reported a history of skin cancer at baseline, there was a reduced odds of NMSC among regular NSAID users for ≥ 5 years (OR 0.82, 95% CI: 0.69–0.98) and b 5 years (OR 0.82, 95% CI: 0.70–0.95) when compared to no/low-users (Pinteraction = 0.04) (Table 3). Inconsistent NSAID use was not associated with statistically significant reduced odds of NMSC (OR 0.90, 95% CI: 0.76–1.07).
11
Discussion Overall, there was no association between NSAID use and NMSC risk in 54,728 post-menopausal women. However, in 5325 women with prior skin cancer history, regular NSAID use was associated with significantly decreased odds of NMSC by 18% relative to no/low-users. To our knowledge, the WHI OS is the first study to demonstrate the association between regular NSAID use and NMSC in women with a history of skin cancer. Our results are consistent with two case–control studies (Johannesdottir et al., 2012; Torti et al., 2011) and one randomized-controlled trial (Elmets et al., 2010). A double-blind, placebo-controlled trial of 240 subjects with actinic keratosis showed a 59% reduction in NMSC in the celecoxib arm versus placebo after 11 months (Elmets et al., 2010), suggesting that our follow-up of 6.9 years is sufficient to observe the preventative effect of NSAIDs on NMSC. In a Denmark population-based case–control study, men and women with NSAID ever-use (N2 prescriptions) compared with nonuse (b2 prescriptions) had a decreased risk of NMSC, especially for longterm (N7 years) and high-intensity use (Johannesdottir et al., 2012). Another case–control study found that regular use of full dose (N200 mg aspirin), but not low dose (b200 mg) NSAIDs was associated reduced odds of NMSC (Butler et al., 2005). Interestingly, we did not see a significant dose–response relationship with longer duration of NSAID use. One explanation for our results may be differences in study design and definitions of “regular” NSAID use. Further studies are needed to evaluate the ideal duration and dosage of NSAID use for NMSC chemoprevention as this has implications for limiting side effects from long-term use. Our findings differed from those of a case–control study (Asgari et al., 2010) and cohort studies (Friis et al., 2003; Grau et al., 2006; Jeter et al., 2012). The Nurses' Health Study cohort found no association between NSAIDs and NMSC risk; however, women with a history of skin cancer were excluded (Jeter et al., 2012). Likewise a Kaiser case–control study showed no reduction in risk but also excluded subjects with prior skin cancer history in their controls (Asgari et al., 2010). Other explanations for the inconsistency with our findings include the evaluation of NSAIDs on the risk of SCC alone rather than all NMSCs (Asgari et al., 2010), smaller sample size (Asgari et al., 2010; Grau et al., 2006), differing definitions of regular NSAID use (Asgari et al., 2010; Grau et al., 2006; Jeter et al., 2012) and exposure measurements relying strictly on recall without medication validation (Jeter et al., 2012). We believe our findings that NSAIDs are associated with decreased odds of NMSC only in women with a prior history of skin cancer are consistent with studies of other malignancies in high-risk populations (Ferrandez et al., 2012). Specifically, chronic inflammation correlates with increased risk of developing cancer in the affected organ and can be caused by ultraviolet radiation. Subjects with prior skin cancer have had greater ultraviolet-B exposure (Chang et al., 2010) and have higher epidermal COX2 expression and NF-kB activation, resulting in inflammation and cancer pathogenesis. These compounds are inhibited by NSAIDs, making this population more amenable to chemoprevention (Yamamoto and Gaynor, 2001). Our findings suggest that NSAIDs may be more effective in secondary prevention among high-risk patients rather than in primary prevention of NMSC, much like studies showing greater benefits of antihypertensives in those with higher versus lower categories of hypertension (Arguedas et al., 2009). The strengths of our study include the prospective design, large number of cases and long follow-up period as well as detailed information on key skin cancer risk factors. In addition, NSAID use was validated with pill bottle labels and/or prescription records, which may not have been done in other studies (Jeter et al., 2012). Limitations include selfreporting of NMSC (without medical record adjudication), although the validity of patient self-reporting of skin cancer is high (Ming et al., 2004) and lack of information on NMSC subtype (BCC vs SCC), which may have yielded information on whether NSAIDs have a specific effect based on subtype, as demonstrated previously (Johannesdottir et al.,
12
A. Wysong et al. / Preventive Medicine 69 (2014) 8–12
2012; Torti et al., 2011). Finally, we did not stratify the analysis by NSAID type or dosage among women with a history of skin cancer and cannot comment on whether ASA versus non-ASA NSAIDS or high versus low dosage NSAIDs were driving the protective effect. Conclusion In conclusion, our results indicate that although NSAID use was not associated with NMSC risk overall, postmenopausal Caucasian women with a history of skin cancer who regularly used NSAIDs had significantly lower odds of future NMSC. Future studies on potential chemopreventative effects of NSAIDs should be focused on subjects with prior history of NMSC. Conflict of Interest statement The authors declare that they have no conflict of interest.
Acknowledgments The authors would like to thank Kristopher Kapphahn and Haley Hedlin for their help with the statistical analysis. References Arguedas, J.A., Perez, M.I., Wright, J.M., 2009. Treatment blood pressure targets for hypertension. Cochrane Database Syst. Rev. CD004349 (Online). Asgari, M.M., Chren, M.M., Warton, E.M., Friedman, G.D., White, E., 2010. Association between nonsteroidal anti-inflammatory drug use and cutaneous squamous cell carcinoma. Arch. Dermatol. 146, 388–395. Butler, G.J., Neale, R., Green, A.C., Pandeya, N., Whiteman, D.C., 2005. Nonsteroidal anti-inflammatory drugs and the risk of actinic keratoses and squamous cell cancers of the skin. J. Am. Acad. Dermatol. 53, 966–972. Chan, A.T., Ogino, S., Giovannucci, E.L., Fuchs, C.S., 2011. Inflammatory markers are associated with risk of colorectal cancer and chemopreventive response to anti-inflammatory drugs. Gastroenterology 140, 799–808 (quiz e11). Chang, N.B., Feng, R., Gao, Z., Gao, W., 2010. Skin cancer incidence is highly associated with ultraviolet-B radiation history. Int. J. Hyg. Environ. Health 213, 359–368.
Elmets, C.A., Viner, J.L., Pentland, A.P., et al., 2010. Chemoprevention of nonmelanoma skin cancer with celecoxib: a randomized, double-blind, placebo-controlled trial. J. Natl. Cancer Inst. 102, 1835–1844. Ferrandez, A., Piazuelo, E., Castells, A., 2012. Aspirin and the prevention of colorectal cancer. Best Pract. Res. Clin. Gastroenterol. 26, 185–195. Friis, S., Sorensen, H.T., McLaughlin, J.K., Johnsen, S.P., Blot, W.J., Olsen, J.H., 2003. A population-based cohort study of the risk of colorectal and other cancers among users of low-dose aspirin. Br. J. Cancer 88, 684–688. Gamba, C.A., Swetter, S.M., Stefanick, M.L., et al., 2013. Aspirin is associated with lower melanoma risk among postmenopausal Caucasian women: the Women's Health Initiative. Cancer 119 (8), 1562–1569. Grau, M.V., Baron, J.A., Langholz, B., et al., 2006. Effect of NSAIDs on the recurrence of nonmelanoma skin cancer. Int. J. Cancer 119, 682–686. Hays, J., Hunt, J.R., Hubbell, F.A., et al., 2003. The Women's Health Initiative recruitment methods and results. Ann. Epidemiol. 13, S18–S77. Higashi, Y., Kanekura, T., Kanzaki, T., 2000. Enhanced expression of cyclooxygenase (COX)-2 in human skin epidermal cancer cells: evidence for growth suppression by inhibiting COX-2 expression. Int. J. Cancer 86, 667–671. Jeter, J.M., Han, J., Martinez, M.E., Alberts, D.S., Qureshi, A.A., Feskanich, D., 2012. Non-steroidal anti-inflammatory drugs, acetaminophen, and risk of skin cancer in the Nurses' Health Study. Cancer Causes Control 23, 1451–1461. Johannesdottir, S.A., Chang, E.T., Mehnert, F., Schmidt, M., Olesen, A.B., Sorensen, H.T., 2012. Nonsteroidal anti-inflammatory drugs and the risk of skin cancer: a population-based case–control study. Cancer 118, 4768–4776. John Chen, G., Yelverton, C.B., Polisetty, S.S., et al., 2006. Treatment patterns and cost of nonmelanoma skin cancer management. Dermatol. Surg. 32, 1266–1271. Ming, M.E., Levy, R.M., Hoffstad, O.J., Filip, J., Gimotty, P.A., Margolis, D.J., 2004. Validity of patient self-reported history of skin cancer. Arch. Dermatol. 140, 730–735. National Institutes of Health, Women's Health Initiative, Form 44-Current Medications, 1998. Design of the Women's Health Initiative clinical trial and observational study. The Women's Health Initiative Study Group. Control. Clin. Trials 19, 61–109. Stern, R.S., 2010. Prevalence of a history of skin cancer in 2007: results of an incidence-based model. Arch. Dermatol. 146, 279–282. Tang, J.Y., Aszterbaum, M., Athar, M., et al., 2010. Basal cell carcinoma chemoprevention with nonsteroidal anti-inflammatory drugs in genetically predisposed PTCH1+/− humans and mice. Cancer Prev. Res. (Phila.) 3, 25–34. Torti, D.C., Christensen, B.C., Storm, C.A., et al., 2011. Analgesic and nonsteroidal anti-inflammatory use in relation to nonmelanoma skin cancer: a population-based case–control study. J. Am. Acad. Dermatol. 65, 304–312. Yamamoto, Y., Gaynor, R.B., 2001. Therapeutic potential of inhibition of the NF-kappaB pathway in the treatment of inflammation and cancer. J. Clin. Invest. 107, 135–142.
